Snowflake26 дней назад

Senior Software Engineer - Cortex AI - FDE

Зарплата не указана

РЫНОК

15 044 ₽медиана по профессии

Backend Developer · 82 вакансий с указанной зарплатой

5 092половина предложений: 11 885–20 1031,2 млн

Работодатель не указал зарплату — сравните с рынком сами.

Полная занятостьУдалёнка

Обязанности

01Architect Agentic Runtimes: Build and scale the orchestration engines that execute complex agentic workflows, ensuring low-latency tool execution and robust state management.
02Scale Context Engineering Infra: Design high-performance systems for RAG (Retrieval-Augmented Generation), including vector database integration, scalable and efficient search indexing, query processing, and result ranking, semantic caching, and automated metadata extraction.
03Build the "Evals Engine": Develop the automated infrastructure required to run massive-scale golden set simulations, error analysis pipelines, and "hillclimbing" experiments.
04Productionize AI Workflows: Collaborate with the modeling team to take raw LLM capabilities and turn them into hardened, multi-tenant microservices with strict guardrails and observability.
05Optimize Performance & Cost: Direct the infra strategy for model routing, prompt caching, and token optimization to ensure Snowflake’s AI features are the most efficient in the industry.

01Education: Bachelor’s degree in Computer Science or a related technical field.
02Experience: 7+ years of experience building distributed systems, high-throughput APIs, or backend infrastructure for AI/ML products.
03Technical Stack: Deep proficiency in Go or Java (for systems) and Python (for AI orchestration).
04Systems Thinking: Strong understanding of database internals, distributed state management, and cloud-native architecture (Kubernetes, FoundationDB, etc.).
05Domain Expertise: Familiarity with the "plumbing" of AI: vector indices, agent platforms, and building scalable data pipelines.
06Experience in a customer-facing technical role — you have explained a hard failure to a frustrated external audience and been believed, you can produce both the internal analysis and the customer-safe version.
07Product instinct — you can judge whether one customer's problem is bespoke or a platform gap worth fixing for everyone. This is the core judgment call.
08Cross-layer debugging — tracing a request across services and root-causing in unfamiliar code from logs and telemetry, not guesswork.
09Eval frameworks for LLM/agent systems — defining quality metrics and using evals to improve quality systematically over time.
10Comfort with ambiguity on open-ended, externally-driven problems.

01For jobs located in the United States, please visit the job posting on the Snowflake Careers Site for salary and benefits information: careers.snowflake.com http://careers.snowflake.com